This is a study to investigate the diagnostic and prognostic value of MRI in spinal cord injury. We performed this prospective study on sixty two patients of acute spinal trauma. We evaluated the epidemiology of spinal trauma & various traumatic findings by MRI. MRI findings were correlated with clinical findings at admission & discharge according to ASIA impairment scale. Four types of MR signal patterns were seen in association with spinal cord injury-cord edema / non haemmorhagic cord contusion (CC), severe cord compression (SCC), cord hemorrhage (CH) and epidural heamatoma (EH). Isolated lesion of cord contusion was found in 40%. All other MR signal patterns were found to be in combination. In cord contusion we further subdivided the group into contusion of size < 3 cm and contusion of size > 3 cm to evaluate any significance of length of cord contusion. In cord heammorhage involving >1cm of the cord, focus was said to be sizable. On bivariate analysis, there was a definitive correlation of cord contusion (CC) involving <3cm & > 3cm of cord with sensory outcome. In >3cm, chances of improvement was 5.75 times lesser than in patients with CC involving <3cm of cord (odds ratio = 5.75 (95% CI: 0.95, 36), Fisher's exact p = 0.0427 (p<.05). In severe cord compression (SCC) the risk of poor outcome was more (odds ratio 4.3 and p=0.149) however was not statistically significant. It was noted that the patients in which epidural hematoma (EH) was present, no improvement was seen, however, by statistical analysis it was not a risk factor and was not related with the outcome (odds ratio – 0.5 and p = 0.22). Presence of cord oedema / non haemorrhagic contusion was not associated with poor outcome (odds ratio 0.25 and p=0.178). On multiple logistic regression / multivariate analysis for estimating prognosis, sizable focus of haemorrhage was most consistently associated with poor outcome (odds ratio −6.73 and p= 0.32) however it was not statistically significant. The risk of retaining a complete cord injury at the time of follow up for patients who initially had significant haemorrhage in cord was more than 6 fold with patients without initial haemorrhage (odds ratio 6.97 and p= .0047). Besides being helpful in diagnosis, MRI findings may serve as a prognostic indicator for clinical, neurological and functional outcome in acute spinal trauma patients